This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mohr, D. Articles by Marmar, C. R. Search for Related Content PubMed PubMed Citation Articles by Mohr, D. Articles by Marmar, C. R. Related Collections Somatoform Sleep and Biological Rhythms Stress and Coping

The Mediating Effects of Sleep in the Relationship Between Traumatic Stress and Health Symptoms in Urban Police Officers

From the University of California, San Francisco; and the Department of Veterans Affairs Medical Center, San Francisco, California.

Address reprint requests to: David C. Mohr, PhD, VA Medical Center, 4150 Clement Street (116A), San Francisco, CA 94121. Email: dmohr{at}itsa.ucsf.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
OBJECTIVE: Posttraumatic stress symptoms have been associated with increased health problems across numerous studies. Sleep disruption, one of the principal symptoms resulting from traumatic stress, has also been shown to produce health problems. This study explored the hypothesis that the relationship between posttraumatic stress symptoms and health is mediated by sleep problems.

METHOD: A sample of 741 police officers were administered measures of traumatic stress symptoms, sleep, health functioning, and somatic symptoms.

RESULTS: Traumatic stress symptoms were significantly related to both somatic symptoms (R² = 0.18, p < .001) and health functioning (R² = 0.02, p < .01). The relationship between somatic symptoms and traumatic stress symptoms was partially mediated by sleep (p < .001). The relationship between traumatic stress symptoms and health functioning was fully mediated by sleep.

CONCLUSIONS: Although design characteristics, such as cross-sectional sampling, limit the inferences that can be drawn, these findings suggest that sleep may serve as an important mediator between traumatic stress and somatic symptoms.

Key Words: posttraumatic stress, • sleep, • health, • somatic symptoms.

Abbreviations: CIHQ = Critical Incident History Questionnaire;; MAST = Michigan Alcoholism Screening Test;; MS-CV = Mississippi Scale – Civilian Version;; PSQI = Pittsburgh Sleep Quality Index;; PTSD = posttraumatic stress disorder;; SCL-90 Som = Symptom Checklist-90, Somatic Symptom Scale;; SF-12 PCS = Short-Form 12, Physical Composite Score.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
The development of posttraumatic stress symptoms has been associated with increased physical health problems across numerous studies (1, 2). However, the mechanisms that account for this association remain poorly understood. There is substantial evidence from the general population that sleep problems are associated with poor health. Large community studies have reported that sleep disturbances are associated with increased rates of functional impairment, medical disability, and treatment utilization (3, 4) . Sleep difficulties are associated with increased use of medical services even among individuals with no previous health problems (5).

The relationship of sleep disturbances to health problems in the general population raises the possibility that sleep disturbances may explain health problems in individuals with posttraumatic stress disorder (PTSD). Sleep difficulties are one of the most frequently noted symptoms in individuals suffering from PTSD. Individuals with PTSD have reported a wide variety of sleep problems including recurrent nightmares, motor dyscontrol during sleep, self-inflicted physical injuries during sleep, and night terrors (6, 7). These sleep disturbances have been linked specifically to traumatic stress symptoms (8). Several polysomnographic studies have found changes in the architecture of sleep in PTSD (6, 9, 10) , although these findings have not always been replicated (11).

Given that sleep problems are extremely common in individuals experiencing posttraumatic stress symptoms, they may provide an important link between posttraumatic stress and health. In a recent study of 167 rape victims, trauma-related sleep disturbances predicted unique variance in physical health symptoms after controlling for other PTSD and depression symptoms (12). However, no study to date has examined whether or not the effects of traumatic stress on health symptoms are mediated by sleep. Such mediation effects would be potentially useful in suggesting potential therapeutic strategies in reducing the adverse effects of trauma on health.

The current study examines the relationship between posttraumatic stress symptoms, sleep problems, and health functioning and somatic symptoms in urban American police officers. We hypothesized that posttraumatic stress symptoms would be related to health functioning and somatic symptoms. We further hypothesized that these relationships would be explained partially or fully by sleep disturbances.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Participants
Police officers were recruited from the police departments of New York, New York, and Oakland and San Jose, California. Potential participants were selected from departmental personnel rosters without identifying specific sampling strata other than the selection for a proportionally higher pool of minority and female officers. Officers were recruited via separate letters of invitation from the project team, the police commissioner, and a representative of the police union. The study team contacted persons who mailed back an affirmative postcard. After providing informed consent, participants were mailed a survey booklet containing study measures, which they completed and returned by mail directly to the research group. Participants were compensated $100 for their efforts.

Measures
Health function was measured using the Physical Composite Score (PCS) of the Medical Outcomes Study 12-Item Short Form Health Survey (SF-12) (13). The SF-12 PCS is a self-report measure that assesses the degree to which physical symptoms interfere with activities. Higher scores on the SF-12 PCS reflect better health functioning. The observed standardized Cronbach of the unweighted items in this administration was = 0.85.

Somatic symptoms were measured using the Somatization Subscale of the Symptom Checklist-90-R (SCL-Som) (14). This scale measures the intensity of self-reported somatic complaints during the previous week including headache, faintness/dizziness, pains in the heart or chest, pains in lower back, muscle soreness, numbness/tingling, and heavy feelings in arms or legs. Items related to sleep were removed. The observed standardized Cronbach in this administration was = 0.84.

Demographic factors were acquired, including age, gender, family income, education, ethnicity, sexual orientation, and relationship status.

Alcohol abuse was assessed using the Michigan Alcoholism Screening Test (MAST) (15). The observed standardized Cronbach of the unweighted items in this administration was = 0.64.

Duty-related trauma history was assessed using the Critical Incident History Questionnaire (CIHQ) (16, 17). The CIHQ measures the number of times the respondent was exposed to each of 34 police-related critical incidents (eg, being present when a fellow officer was killed, being shot at, making a mistake that led to the serious injury or death of a bystander) and each officer’s ratings of how difficult each incident would be for a fellow officer to cope with. In the present study, the total cumulative exposure score was derived by summing across all items the frequency by weighted difficulty of exposures to critical incidents. Because this is a checklist evaluating the occurrence of events in the participants’ environment, measures of internal reliability are not relevant. However, the validity of the CIHQ is supported by high correlations with other measures of traumatic symptoms as well as years of police service (17).

Traumatic stress symptoms secondary to critical incidents during police service were evaluated by the Mississippi Scale–Civilian Version (MS-CV), which is a modification of Mississippi Scale for Combat-Related Posttraumatic Stress Disorder (18, 19). In scoring the MS-CV, we excluded items that pertained either to sleep or to physical health. The observed standardized Cronbach for this administration was = 0.86.

Subjective sleep problems were evaluated using the Pittsburgh Sleep Quality Index (PSQI) (20). The PSQI is a 19-item self-report questionnaire that assesses sleep quality and disturbances, validated by polysomnography, that have occurred during the past month. The 19 individual items yield seven component scores: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction. Scores on the seven dimensions are summed to form a single global score. The observed standardized Cronbach of the unweighted Likert scale items was = 0.71. This reliability report excludes items assessing absolute sleep latency and duration in minutes and hours.

Statistical Analyses
We hypothesized that sleep problems would mediate or explain the relationship between traumatic stress symptoms and both health functioning and somatic symptoms. Mediation analyses are used to examine if a mediator variable can explain how or why a preceding variable affects an outcome (21). All three variables must be significantly related with one another. Logistic regression analyses were conducted to establish these relationships. Each set of mediational analyses contained four regression analyses to obtain coefficients of determination (R²). 1) Trauma symptoms, residualized for demographics and other control variables, must be significantly related to the PSQI. 2) The PSQI, residualized for demographics and other control variables, must be related to each of the dependent health variables, SF-12 PCS, or SCL-Som. 3) Trauma symptoms, residualized for demographics and other control variables, must be significantly related to each of the dependent variables. 4) The relationship between trauma and each dependent health variable, after being residualized for the PSQI, must be either nonsignificant or significantly less than the unresidualized relationship. If it is nonsignificant, PSQI can be said to "fully mediate" the relationship between trauma and the dependent health variables. If it is significant, but there is also a significant reduction in the strength of the relationship, the PSQI can be said to "partially mediate" the relationship between trauma and the dependent health variables.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Participants
Of the 1200 officers who indicated a willingness to participate and were mailed questionnaire booklets, a total of 790 (66%) returned self-report questionnaire booklets. Forty-nine returned questionnaire booklets that were unusable. Of the 741 usable booklets, 713 officers completed the SF-12 PCS subscale and 733 officers completed the SCL-Som. Missing independent variables were imputed using the expectation maximization method described by Little and Rubin (22).

The police officers were an average of 37.1 (SD = 6.9) years old and 21.4% were women. All had at least a high school education and 68.2% had completed an associate’s, bachelor’s, or graduate degree. They were a diverse sample, with 44.0% Caucasian, 22.3% African-American, 26.3% Hispanic, and 7.4% other ethnicities. Most (69.1%) were married.

Preliminary Analyses
We were concerned that police officers would be healthier than the general population. Furthermore, examination of our dependent variables showed that they were not normally distributed. We therefore decided to dichotomize the variables. Normative samples recommend a SCL-Som cutoff of 0.70 or higher as defining high somatic symptoms (14). This resulted in 17.3% of the sample being classified as showing significant somatic symptoms. The symptomatic group had a mean SCL-Som score of 1.14 (SD = 0.41) and the nonsymptomatic group had a mean score of 0.23 (SD = 0.19).

Because similar cutoffs are unavailable for the SF-12 PCS, we used the percentage of SCL-Som somatic versus nonsomatic subjects to dichotomize the SF-12 PCS. Dividing the sample into the 17.3% with the lowest health function and the 82.7% with the best health function resulted in SF-12 PCS cutoff of 47.8. The low health function group had a mean SF-12 PCS score of 39.2 (SD = 6.7) and the high health function group had a mean score of 55.3 (SD = 3.11).

All analyses control for four potentially confounding variables by including them in the first step of each hierarchical regression analysis. These variables included age, alcohol abuse, Hispanic ethnicity, and exposure to police-related critical incidents (CIHQ). A dichotomized variable for Hispanic ethnicity was entered because Hispanic ethnicity has been shown to be a risk factor for PTSD (23).

Traumatic stress symptoms and sleep.
The relationship between traumatic stress symptoms and sleep problems is important for both mediational analyses. After accounting for the four potentially confounding variables (R² = 0.024, p < .001), traumatic stress symptoms were significant predictors of the PSQI (R² = 0.093, p < .001).

Health functioning.
The results of this analysis are shown in Table 1. Using SF-12 PCS as the dependent variable, the four potentially confounding variables explained a significant increment in explained variance of poor health functioning (R² = 0.079, ² = 34.0, p < .001). Traumatic stress symptoms, added in the second step, accounted for a further significant increment in variance (R² = 0.019, ² = 8.6, p < .01). Sleep problems, added in the third step, also accounted for a significant portion of the remaining variance (R² = 0.061, ² = 27.7, p < .001). Reversing the order of entry for the last two steps, sleep problems showed a significant association with the SF-12 PCS (R² = 0.077, ² = 34.9, p < .001). However, when traumatic stress symptoms were entered last, they were not independently associated with the SF-12 PCS (R² = 0.003, ² = 0.14, p = .23). Thus, sleep problems meet criteria as a full mediator of the relationship between traumatic stress symptoms and health functioning.

View larger version (14K): [in this window] [in a new window]   Fig. 1. Model 1: relationship between PTSD and somatic symptoms. Model 2: sleep as a mediator of the relationship between PTSD and somatic symptoms.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Our hypothesis that the relationship between traumatic stress symptoms and health would be mediated by subjective sleep quality was generally supported. Subjective sleep partially mediated the relationship between traumatic stress and somatic symptoms. The reduction of 45.4% in the relationship is substantial.

Subjective sleep quality fully mediated the relationship between traumatic stress and health functioning. However, although the relationship between traumatic stress and health functioning met the statistical criteria for significance, the effect size for the unmediated relationship was small (R² = 0.02) and of questionable clinical significance. This small effect size was likely due to basement effects on our measure of health functioning, the SF-12 PCS, in this population. The SF-12 PCS primarily assesses functional impairment such as difficulty in ambulation. Thus, although some of these officers experienced moderate-to-high levels of duty-related traumatic stress, as well as somatic symptoms associated with that stress, they did not experience severe enough problems in health functioning to be disabled for active duty service. This likely restricted the range on our measure of health function.

Our findings extend the work of others who have identified relationships between trauma, health, and sleep difficulties. Our findings are consistent with those of Clum and colleagues (12) who showed in a sample of rape victims that sleep problems were significantly related to health symptoms independent of symptoms of traumatic stress. Similar findings have been reported by Kiecolt-Glaser (24) for chronic nontraumatic stress among caregivers of Alzheimer’s. Our findings extend those of Clum and Kiecolt-Glaser by suggesting that sleep is not only an independent predictor of health symptoms but also serves as a mediator between trauma and health. Although the mechanism by which sleep might mediate the relationship between trauma and health is unclear, sleep-related changes in immunity likely play a role. Changes in immune activity following a natural disaster have been shown to be mediated by sleep problems (25).

Caution is required in the interpretation of our findings because of design and measurement limitations. This is a study of subjectively perceived rather than objectively rated somatic symptoms and health functioning. The relationship between self-reported and physician-rated physical health among patients with PTSD is variable (26). Similarly, subjective reports of sleep disturbance are not always highly correlated with polysomnographic sleep data (11). A further limitation of our study is the cross-sectional design. We have assumed that sleep problems are a result of traumatic stress and a cause of health problems. However, we cannot rule out the possibility that sleep problems aggravate traumatic stress symptoms. Although a reciprocal relationship is possible, a large body of work supports the notion that traumatic stress causes sleep disturbance (9, 10, 27, 28).

In spite of the methodological limitations, this study extends the existing literature on the interrelationship between a cluster of problems associated with traumatic stress. Our finding that sleep disturbances mediate the relationship between traumatic stress symptoms and health variables suggests that improving sleep may modulate the adverse effects of traumatic stress on health. Future research evaluating behavioral and pharmacologic treatments for sleep disorders associated with traumatic stress may prove to be invaluable to the secondary prevention of health problems in the millions of people who experience traumatic stress each year.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
This research was supported by National Institute of Mental Health Grant MH56350-01A1 (C.R.M.).

Received for publication March 1, 2002.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 

1. Friedman MJ, Schnurr PP. The relationship between trauma, posttraumatic stress disorder, and physical health. In: Friedman MJ, Charney DS, Deutch AY, editors. Neurobiological and clinical consequences of stress: from normal adaptation to posttraumatic stress disorder. Philadelphia: Lippincott-Raven Publishers; 1995. p. 507–24.
2. Deykin EY, Keane TM, Kaloupek D, Fincke G, Rothendler J, Siegfried M, Creamer K. Posttraumatic stress disorder and the use of health services. Psychosom Med 2001; 63: 835–41.[Abstract/Free Full Text]
3. Newman A, Enright P, Manolio T, Haponik E, Wahl P. Sleep disturbance, psychosocial correlates, and cardiovascular disease in 5201 older adults: the Cardiovascular Health Study. J Am Geriatr Soc 1997; 45: 1–7.[Medline]
4. Simon GE, VonKorff M. Prevalence, burden, and treatment of insomnia in primary care. Am J Psychiatry 1997; 154: 1417–23.[Abstract]
5. Weissman MM, Greenwald S, Nino-Murcia G, Dement WC. The morbidity of insomnia uncomplicated by psychiatric disorders. Gen Hosp Psychiatry 1997; 19: 245–50.[CrossRef][Medline]
6. Mellman TA, Kulick-Bell R, Ashlock LE, Nolan B. Sleep events among veterans with combat-related posttraumatic stress disorder. Am J Psychiatry 1995; 152: 110–5.[Abstract/Free Full Text]
7. Neylan TC, Marmar CR, Metzler TJ, Weiss DS, Zatzick DF, Delucchi KL, Wu RM, Schoenfeld FB. Sleep disturbances in the Vietnam generation: findings from a nationally representative sample of male Vietnam veterans. Am J Psychiatry 1998; 155: 929–33.[Abstract/Free Full Text]
8. Neylan TC, Metzler TJ, Best SR, Weiss DS, Fagan JA, Liberman A, Rogers C, Vedantham K, Brunet A, Lipsy TL, Marmar CL. Critical incident exposure and sleep quality in police officers. Psychosom Med 2002; 64: 345–52.[Abstract/Free Full Text]
9. Lavie P, Hefez A, Halperin G, Enoch D. Long-term effects of traumatic war-related events on sleep. Am J Psychiatry 1979; 136: 175–8.[Abstract/Free Full Text]
10. Ross RJ, Ball WA, Dinges DF, Kribbs NB, Morrison AR, Silver SM, Mulvaney FD. Rapid eye movement sleep disturbance in posttraumatic stress disorder. Biol Psychiatry 1994; 35: 195–202.[CrossRef][Medline]
11. Hurwitz TD, Mahowald MW, Kuskowski M, Engdahl BE. Polysomnographic sleep is not clinically impaired in Vietnam combat veterans with chronic posttraumatic stress disorder. Biol Psychiatry 1998; 44: 1066–73.[CrossRef][Medline]
12. Clum GA, Nishith P, Resick PA. Trauma-related sleep disturbance and self-reported physical health symptoms in treatment-seeking female rape victims. J Nerv Ment Dis 2001; 189: 618–22.[CrossRef][Medline]
13. Ware JE, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey – Construction of Scales and Preliminary Tests of Reliability and Validity. Med Care 1996; 34: 220–33.[CrossRef][Medline]
14. Derogatis LR. SCL-90-R administration, scoring, and procedures manual. Minneapolis: National Computer Systems, Inc.; 1994.
15. Selzer ML, Vinokur A, van Rooijen L. A self-administered Short Michigan Alcoholism Screening Test (SMAST). J Stud Alcohol 1975; 36: 117–26.[Medline]
16. Brunet A, Weiss D, Best SR, Liberman A, Fagan J, Marmar CR. Assessing recurring traumatic exposure: the Critical Incident History Questionnaire, Washington DC: International Society for Traumatic Stress Studies; 1998. p. 49.
17. Weiss D, Brunet A, Best S, Metzler T, Liberman A, Rogers C, Neylan T, Fagan J, Marmar C. The Critical Incident History Questionnaire: a measure of cumulative exposure to critical incidents. Manuscript submitted for publication, 2000.
18. Keane TM, Fairbank JA, Caddell JM, Zimering RT, Taylor KL, Mora C. Clinical evaluation of a measure to assess combat exposure. Psychol Assess 1989; 1: 53–5.
19. Marmar CR, Weiss DS, Metzler TJ, Ronfeldt HM, Foreman C. Stress responses of emergency services personnel to the Loma Prieta earthquake Interstate 880 freeway collapse and control traumatic incidents. J Trauma Stress 1996; 9: 63–85.[CrossRef][Medline]
20. Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 1989; 28: 193–213.[CrossRef][Medline]
21. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986; 51: 1173–82.[CrossRef][Medline]
22. Little RJA, Rubin DB. Statistical analysis with missing data. New York: Wiley; 1987.
23. Pole N, Best S, Weiss D, Metzler T, Liberman A, Fagan J, Marmar C. Effects of gender and ethnicity on duty-related posttraumatic stress symptoms among urban police officers. J Nerv Ment Dis 2001; 189: 442–8.[CrossRef][Medline]
24. Kiecolt-Glaser JK, Dura JR, Speicher CE, Trask OJ, Glaser R. Spousal caregivers of dementia victims: longitudinal changes in immunity and health. Psychosom Med 1991; 53: 345–62.[Abstract/Free Full Text]
25. Ironson G, Wynings C, Schneiderman N, Baum A, Rodriguez M, Green D, Benight C, Antoni M, LaPerriere A, Huang H, Klimas N, Fletcher M. Posttraumatic stress symptoms, intrusive thoughts, loss, and immune function after Hurricane Andrew. Psychosom Med 1997; 59: 128–41.[Abstract/Free Full Text]
26. Beckham JC, Moore SD, Feldman ME, Hertzberg MA, Kirby AC, Fairbank JA. Health status, somatization, and severity of posttraumatic stress disorder in Vietnam combat veterans with posttraumatic stress disorder. Am J Psychiatry 1998; 155: 1565–9.[Abstract/Free Full Text]
27. Fabsitz RR, Sholinsky P, Goldberg J. Correlates of sleep problems among men: the Vietnam Era Twin Registry. J Sleep Res 1997; 6: 50–6.[CrossRef][Medline]
28. Mellman TA, Nolan B, Hebding J, Kulick-Bell R, Dominguez R. A polysomnographic comparison of veterans with combat-related PTSD, depressed men, and non-ill controls. Sleep 1997; 20; 46–51.[Medline]

This article has been cited by other articles:

				C. R. MARMAR, S. E. MCCASLIN, T. J. METZLER, S. BEST, D. S. WEISS, J. FAGAN, A. LIBERMAN, N. POLE, C. OTTE, R. YEHUDA, et al. Predictors of Posttraumatic Stress in Police and Other First Responders Ann. N.Y. Acad. Sci., July 1, 2006; 1071(1): 1 - 18. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mohr, D. Articles by Marmar, C. R. Search for Related Content PubMed PubMed Citation Articles by Mohr, D. Articles by Marmar, C. R. Related Collections Somatoform Sleep and Biological Rhythms Stress and Coping